A Pipeline for Developing Population-Specific Axonal-Embedded Brain Finite Element Models

Introduction: : Traumatic brain injury (TBI) is a significant health burden worldwide. Brain finite element (FE) models, which can offer insights into tissue responses during biomechanical TBI events, can serve as brain monitoring tools and accelerate technical innovations in head protective equipment—ultimately reducing TBI incidence. These models also help investigate how subject- or population-specific brain structural variations and other factors affect biomechanical responses in trauma events. Incorporating axonal tractography using DTI further empowers these models to account for the effect of axonal tract distributions on brain responses in trauma events [1, 2]. However, many existing brain FE models are based on a single subject and/or lack critical features like sulci and gyri. While some efforts have been made to morph these generic models to subject-specific brain spaces [3, 4], they are often limited to morphing external or gross features, missing internal details, or failing to incorporate subject- or population-specific tractography patterns. Moreover, population-averaged templates integrating both volumetric MRI and DTI-derived axonal fiber data for specific demographics are missing. Without such templates, FE models are often either single-subject-based or rely on morphed generic shapes, limiting their utility in population-level studies.
As a step forward, we developed a pipeline that generates population-specific MRI and DTI templates, representing average anatomical features and tractography patterns of a population, and then produces hex-only FE brain meshes and axonal fiber beams directly from these population-specific MRI and DTI data. This pipeline preserves fine anatomical details and embeds a population-averaged axonal fiber tractography template within the volumetric mesh.

Materials and

Methods: : As a demonstration of our pipeline, we developed a young adult male-specific, axonal-embedded brain FE model in this study (ages 20–40). First, we created a T1-weighted MRI template from 500 healthy adult males (20–40 years) as described in [5]. Next, DTI data from the same population were processed using Q-space diffeomorphic reconstruction in DSI-Studio to align the orientation distribution functions (ODFs) within the anatomical space developed in the first step. These ODFs were then averaged to generate the adult male-specific DTI template following the protocol described in [6]. ANTs, FreeSurfer, and DSI-Studio were employed to develop the MRI template, perform anatomical segmentation, and generate the DTI template and tractography, respectively. We then developed a MATLAB-based framework to generate a smooth hexahedral mesh from the MRI-based segmentation, reducing voxelation artifacts at tissue interfaces. The falx and tentorium were manually segmented, and shell elements for them were generated. Tractography was performed on this DTI template using the method described in [6], and axonal fibers were generated as 1D cable elements embedded within the brain mesh using a MATLAB code, following the approach in [1–2]. Material properties were adopted from [2]. LS-DYNA was used to simulate head kinematics derived from a soccer heading study [7]. From simulations, MPS95 and MAS95, representing the 95th percentile of maximum principal strain in brain tissue and the 95th percentile of maximum strain in axonal fibers, respectively, were extracted for 40 brain regions.

Results, Conclusions, and Discussions:: The overall and regional MPS-95 and MAS-95 results for the average kinematics of front-high soccer heading in males are shown in Figure 2 as an example. We extracted these parameters in major brain regions to assess tissue-level deformation under this representative head impact. Among all regions, the brainstem, with its dense axonal architecture, exhibited the highest MPS-95 (0.114), consistent with its known susceptibility to rotational accelerations due to its location near rotational centers [8]. Elevated strain was also observed in cortical and cerebellar areas, with the right cortex and bilateral cerebellar cortices exhibiting MPS-95 values above 0.079, suggesting that peripheral and posterior structures may experience higher deformation under these impact conditions. These findings underscore the importance of cortical folding and brain anatomical details at the brain/CSF boundary in head FE models. In contrast, subcortical gray matter regions, including the putamen, amygdala, and hippocampus, showed the lowest MPS-95 values (0.034–0.046), consistent with their location and relative isolation from rotational centers. The corpus callosum and cerebral white matter, while showing moderate strain (0.050–0.053), remain regions of interest due to their relevance in diffuse axonal injury, as reported in both clinical and simulation studies. Overall, these findings highlight the importance of anatomical and microstructural fidelity in capturing region-specific strain patterns under dynamic loading conditions. By incorporating population-averaged volumetric and fiber architecture into a unified framework, our model provides a more representative foundation for evaluating brain injury risk at the population level. As soccer heading is not an injurious scenario, the observed values are expectedly lower than established MPS-95 and MAS-95 TBI thresholds [1].
This study presents a proof-of-concept application of our pipeline to generate a population-specific axonal-embedded brain FE model using neuroimaging data. The pipeline enables integration of averaged anatomical and tractography features into biomechanically accurate, hex-only brain meshes, preserving anatomical realism while supporting region-specific injury analysis. Ongoing efforts are focused on extending this framework to generate other population-specific models. This platform sets the stage for scalable, demographically tailored modeling of brain biomechanics, with broad applications in TBI research, protective equipment design, and precision neurotrauma risk assessment.

Acknowledgements and/or References (Optional)::

Funding: NSF #2138719)
References :
1. Hajiaghamemar, M., Wu, T., Panzer, M. B., & Margulies, S. S. (2020). Embedded axonal fiber tracts improve finite element model predictions of traumatic brain injury. Biomechanics and modeling in mechanobiology, 19, 1109-1130.
2. Wu, T., Alshareef, A., Giudice, J. S., & Panzer, M. B. (2019). Explicit modeling of white matter axonal fiber tracts in a finite element brain model. Annals of biomedical engineering, 47, 1908-1922.
3. Wu, S., Zhao, W., Wu, Z., McAllister, T., Hu, J., & Ji, S. (2023). Approximating subject-specific brain injury models via scaling based on head–brain morphological relationships. Biomechanics and Modeling in Mechanobiology, 22(1), 159-175.
4. Giudice, J. S., Druzgal, T. J., & Panzer, M. B. (2023). Investigating the effect of brain size on deformation magnitude using subject-specific finite element models. Journal of Neurotrauma, 40(15-16), 1796-1807.
5. Ramzanpour, M., Jafari, B., Smith, J., Allen, J., & Hajiaghamemar, M. (2023). Comprehensive study of sex-based anatomical variations of human brain and development of sex-specific brain templates. Brain Multiphysics, 4, 100077.
6. Yeh, F.-C., Panesar, S., Fernandes, D., Meola, A., Yoshino, M., Fernandez-Miranda, J.C., Vettel, J.M., Verstynen, T., 2018. Population-averaged atlas of the macroscale human structural connectome and its network topology. Neuroimage 178, 57-68
7. Abbasi Ghiri, A., Seidi, M., Wallace, J., Cheever, K., & Memar, M. (2025). Exploring Sex-Based Variations in Head Kinematics During Soccer Heading. Annals of Biomedical Engineering, 1-17.
8. Zimmerman, K. A., Cournoyer, J., Lai, H., Snider, S. B., Fischer, D., Kemp, S., ... & Sharp, D. J. (2023). The biomechanical signature of loss of consciousness: computational modelling of elite athlete head injuries. Brain, 146(7), 3063-3078.